Data Analyst — Energy Retail (Oil & Gas)

About the role
A hands-on analyst for a consumer energy business: pay-as-you-go cooking gas (LPG) delivered to hundreds of thousands of households in fast-growing emerging markets. You own the analysis end to end — framing the question, getting the data, building the numbers and defending them in front of the people who will act on them. You work directly with commercial, operations, finance and management, and with our founders — no account manager in between.
You join our energy data team, alongside a senior data scientist and a data & BI engineer. A meaningful part of the work starts in raw tables, meter logs, payment exports and field-operations files rather than in a tidy warehouse.
We do not require you to come from oil & gas. The business model is much closer to a prepaid mobile operator than to a fuel station: customers top up credit with mobile money, consume in small amounts, and every unit used is recorded by a smart meter. Experience at mobile network operators, telecoms, mobile money or fintech, utilities with prepaid metering, pay-as-you-go solar, and similar high-volume B2C businesses is directly relevant.
What the work looks like
Each customer home has a stove, a smart meter and a gas cylinder owned and maintained by the company. Customers pay only for the gas they use, topping up as little as a single meal’s worth through mobile money, and local technicians replace cylinders before they run out. That produces rich data — consumption, top-ups, cylinder swaps, deliveries, meter events — and commercial questions to match: which customers stay and why, what a customer is worth, where the network makes or loses money, and what to do about it.
Scope
Customer analytics — acquisition, activation, retention and churn, cohort behaviour, revenue and margin per household, lifetime value against the cost of installation
Consumption and top-up behaviour — frequency and size of top-ups, usage patterns, seasonality, early signs that a customer is about to go dormant
Pricing and promotion analysis — the effect of price, top-up incentives and campaigns on usage and margin, measured against proper baselines and control groups
Field operations and unit economics — cylinder swap and delivery efficiency, depot and area performance, technician productivity, cost to serve per household
Forecasting gas demand, cylinder logistics and revenue by area, to support supply, stock and expansion decisions
Anomaly and fraud detection — meter tampering, consumption that does not match cylinder swaps, payment irregularities, unusual field activity
Reporting with single agreed definitions where commercial, operations and finance count customers, revenue or volume differently
Messy data as the normal case: meter connectivity gaps, delayed payment records, re-assigned devices, data from more than one country and system
Requirements
4+ years as a data analyst or in a comparable analytical role, owning your own analyses from question to recommendation
Experience with high-volume B2C transactional data — prepaid mobile or telecoms, mobile money, utilities, pay-as-you-go energy, retail or similar. You know what subscriber-, account- or transaction-level data looks like and where it lies
Commercial analytics in practice: churn and retention, cohort and lifetime-value analysis, segmentation, pricing or campaign evaluation
Strong SQL and Python (pandas or similar), and readiness to work in raw data yourself: unfamiliar schemas, undocumented tables, exports never meant for analysis
Sound statistics for business questions — baselines, control groups, seasonality, telling a real effect from noise
Working with coding agents, seriously. Agents write most of the code here. Bad agent output in analysis fails silently: a join fans out rows, a filter drops half the active customers, a top-up is attributed to the wrong month. We hire for verification habits, not speed — row-count and population checks, reconciliation against a number someone already trusts, a traceable path from raw data to any figure a decision-maker sees
Experience with a BI tool (Power BI, Tableau, Metabase, Superset or similar)
Clear communication: you can explain a result, its limits and what to do about it to a non-technical manager
English sufficient to present findings and defend methodology
Professional proficiency in Russian
Nice to have
Experience in emerging markets or with mobile money data
Experience with IoT, smart-meter or device data
Field operations or logistics analytics — routing, depots, delivery efficiency
Experience with experimentation or A/B testing at scale
What we offer
Ownership of customer and commercial analytics, and room to shape how it is done
Direct access to founders and decision-makers; short feedback loops
Generous AI tooling budget — frontier models are a working expense, not a perk
Hybrid working from our London office
Competitive compensation, discussed individually
How to apply
Share your CV and an answer to any of the questions below, at your choice. Applications without answers are not reviewed.
After a top-up bonus campaign in one city, the number of top-ups rose 20%, but gas consumed per household barely changed and revenue per household fell. Marketing calls it a success, finance calls it a loss. How would you work out what actually happened, what data would you need, and what would you deliver?
Give a real case where two teams or two systems reported different numbers for the same customers, subscribers, revenue or churn. What were the figures, why did they differ, which one did the business treat as truth, and what did you change?
Show a real prompt or task brief you gave an agent for an analysis, and what it produced. Then one case where agent output looked right and was numerically wrong: what was wrong, how you caught it, what check you added.
Which transactional or customer systems have you personally pulled data from — billing or CRM at a telecom operator, mobile money or payment platforms, smart-meter or IoT data, POS, ERP? Pick one and tell us something about its data that only someone who has used it would know.

Employment type
Full-Time

Work arrangement
No